Cloudflare — Services & Pricing

Cloudflare covers a partial, edge-native stack — object storage (R2), serverless (Workers), CDN — with its headline differentiator $0 egress. No VM / block storage / managed-DB / Kubernetes offering.

Cloudflare isn’t a like-for-like alternative to the three majors — it’s an edge-native platform that runs your code and data on its global network, close to users. It deliberately omits the traditional IaaS stack (no VMs, no managed relational DB at scale, no Kubernetes) and instead offers serverless compute (Workers), object storage (R2), KV / D1 (SQLite) / Durable Objects / Queues, and a best-in-class CDN + security layer (WAF, DDoS). Its one structural, hard-to-match advantage is economic: zero egress fees. That’s why it most often appears alongside a major cloud rather than replacing it.

_It doesn’t localize the full reference architecture — it occupies the **edge + delivery + serverless

• object-storage** layers (see the cloud architecture page)._

• Workers — serverless compute that runs at the edge (V8 isolates, near-zero cold start), ideal for latency-sensitive logic, APIs, and front-end back-ends.
• R2 — S3-compatible object storage with no egress fees; the headline TCO play for egress-heavy workloads (media, downloads, data distribution).
• Edge data — KV (global key-value), D1 (SQLite at the edge), Durable Objects (stateful coordination), Queues — a serverless data tier, not a managed RDBMS.
• Network & security — global CDN, WAF, DDoS protection, DNS, Zero Trust / Access — the layer Cloudflare has always led.

• Egress-heavy workloads — R2’s $0 egress vs ~$0.09/GB (AWS), $0.087 (Azure), $0.12 (GCP) is a decisive TCO difference for media, downloads, or multi-cloud data movement.
• Edge-first / latency-sensitive apps — Workers run close to users worldwide; great for APIs, personalization, and JAMstack/serverless front-ends.
• CDN, DNS, and security — best-in-class delivery and protection, often adopted in front of a major-cloud origin.
• As a complement, not a replacement — a common pattern is R2 for storage + Workers at the edge, with compute/managed-DB on a major cloud — capturing the egress savings without giving up IaaS depth.
• Not the fit when: you need VMs, managed relational databases at scale, Kubernetes, or lift-and-shift of traditional enterprise workloads — those belong on AWS/Azure/GCP.

Cost note: R2’s economics are about what’s not charged (egress) as much as the per-GB storage rate. Model total cost including request volume and egress patterns, not storage alone.

Pricing generated from the live cost catalog via scripts/gen_cloud_docs.py.